I am only worried that since OLED will junk such a big area of displays, manufactuerers will be hesitant to deploy it, or will deploy it expensively and with low supply. The good thing is i guess it only takes one company to do it right, and the prices will come crashing down.
From what I hear generally, the main barrier for OLED's is the longevity of the devices. That might be what'll take it a while to get going, because it's still kind of a fundamental engineering problem. Though -- if it's cheap enough, it might not be so bad to junk your monitor after 1 year, and buy another. Though that'd be bad environmentally, etc. Plus it'd be a pain. Ah, well.
Hopefully it will help solve the viewing angle problem with LEDs, that's the reason we still have filiment bulbs for some indicator purposes.
Doesn't look like it -- from the diagram, it's looking very directional (the angles of the backplane reflectors are set up to reflect light from the emitter at the center to a line perpendicular to the panel).
This will be useful for things like traffic lights, brake lights, and home lighting.
One problem: one of the useful things about the cluster-of-bullet-LED's we have now for traffic/brake/etc. lights is an additional fault-tolerance aspect: you could have a substantial number of them malfunctioning and still be useful.
I have actually seen this -- a big truck in front of me had a really annoying one, where 1/3 of them were out, 1/3 on, and 1/3 flickering randomly. Annoying, but it still performed its function.
If you just have one LED chip in the middle, if it goes out, you gotta replace the whole thing.
Oh, it also seems that it'll be harder to get sufficient brightness for the really demanding applications (e.g. a traffic light with the sun behind it).
I've still got a lot of pr0n^H^H^H^H Data on floppy. What will I do if I can't read those?
You mean when you can't read those -- nothing to do with availablility of drives, they're going to rot!. Better get them onto a cd as soon as you can.
... a few years pass...
I've still got a lot of pr0n^H^H^H^H Data on cd. What will I do if I can't read those?
You mean when you can't read those -- nothing to do with availablility of drives, they're going to rot!. Better get them onto a dvd as soon as you can.
... a few years pass...
I've still got a lot of pr0n^H^H^H^H Home videos on dvd. What will I do if I can't read those?
You mean when you can't read those -- nothing to do with availablility of drives, they're going to rot!. Better get them onto a holo-crystal as soon as you can.
... a few years pass...
I've still got a lot of love-slave android personaliti^H^H^H^H Personal memory engram backups on holo-crystal. What will I do if I can't read those?
Ah, but in Soviet Russia, what will you do when your holo-crystal can't read YOU!
You'll have to bring me up-to-date on the history of "insane" boycotting "for any reason" being a "proud American tradition".
I believe he was referring to the matter of free speech.
Are you saying it is RIGHT to boycott...[various examples]
I think he was not saying what is right, but more of what one has a right to do.
...but is the basis of his/her boycott ethical? Is it decent?
Doesn't matter. One has the right to boycott what one wishes. If it is an unfair/unreasonable boycott, then additional voices can be raised against it.
Hey, that'd be a meta-boycott!
Imagine what a beowulf cluster of meta boycotts could... ouch!
I'm no expert, but I thought PRMan also heavily uses shader programs in addition to (or instead of) static textures...
Yes, indeed, programmatic shaders are one of the most important features of the Renerman standard. But I'm betting that, at Pixar, 90% of the time they use texmaps. It just works out to be a lot easier to do it that way -- provided you have an army of artists to generate all those 2D images! (BTW, technically, renderman always uses programmatic shaders; to do texmaps, you use a simple shader that just goes and looks up the color values to set.)
Plus, since it typically isn't doing raytraycing, it needs to do a lot of "shadow map" lookups, to tell whether an object is being directly lit or not. I don't know all the details of that, but it does involve doing a texture-map-like lookup for every light source.
Thanks for the "Renderman" etymology.:-)
You're welcome! I got it from the Advanced Renderman book. There's a few paragraphs on it in the back.
How about 1 DSP per pixel (About 10 million?). I'm sure that would really zip along, if they could sort out the memory access issues inherent in this kind of application.
Aye, there's the rub.
Given that they picked Intel chips over Athlons, and given that they must have carefully compared chip performance on their particular application (i.e. Renderman), that says to me that Renderman is mostly memory-bandwith limited, rather than heavy-math-calculation limited.
I know a little bit about how that program works: it's not a ray-tracer. It does some basic 3d calcs to trace lines from camera to objects, and to subdivide polygons. But to determine the actual color of each pixel, it's mostly a matter of one or a number of texture-map, shadow-map, and other lookups. Each is addressing a small part of a big range of memory -- probably breaks through the cache incessantly. I'd bet that many of the geometric calculations are memory-limited too, due to the absolutely humongous number of objects they put into a given scene (e.g. blue fur).
From what I understand, DSP's are good for 2D image processing -- because the algorithms are fairly standard, and require a lot of signal-processing-like math. For 3D, perhaps the matrix calcs involved in the coordinate transforms could be done in a dsp, but as I described, that's probably not a big enough piece of the puzzle to make it worth it.
To go a little further: dedicated hardware was actually the original goal of Pixar, even before it was split off as its own company. But they noticed that hardware advances were so fast that their designs were getting obsolete fast. (The name "renderman" actually came from a quip by one of their engineers, commenting on how they'd soon be able to design a machine that could fit in a pocket-sized device that you could carry around like a Walkman.) Anyways, they eventually discarded the custom hardware, because their software-only "practice" version was getting quite acceptable performance levels all by itself -- on general-purpose processors.
Fourth, provide incentives to corporations to begin manned space flight outside the scope of NASA oversight. Fifth, turn NASA into a regulatory agency for the purposes of establishing safety guidelines; and a science agency which would fund and oversee pure science activities in space.
I think you've missed something here: what about the excellent astronaut-training facilities/staff/experience that NASA has built up over the decades, as well as the current crop of astronauts? I'd suggest that in addition to your "Fifth" option, add the creation of NATA, the National Astronautics Training Academy. This would consist of just that subset of NASA; its job would be to train the astronauts that companies would hire for space work.
A company may be able to finance the hardware for a space mission, but the astronaut training requires a big, big, long-term investment. Keeping that centralized probably makes sense. This would be a sort of "rent-a-naut" system.:)
I'd also like to see a space elevator persued, but I don't know that we have the tech yet.
Here's a good paper on the subject. It's a 15mb pdf, but worth the download.
In it, a good many of the technical problems are solidly examined, and reasonable solutions are proposed.
The approach presented is to launch an initial spool of very thin cable into geosynchronous orbit. This spool will be some thousands of kg in mass; this won't be *that* much harder than putting up a communications satellite. Then you lower that cable down to earth (and raise spool-unreeling spacecraft up past geosynch. as a counterweight), and you have a sort of "mini" space elevator that can haul up a mere 1200 kg. A series of climbers then ascend, each epoxying on a new layer of cable. Continue for 2 years, and you have a cable that can carry up as much as the shuttle. Continue for 5 years, and you have one that can lift a million kg.
All the solutions to the technical problems will require lots of research/testing to truly overcome, so it'd likely still be decades away, even with full effort. And that's also assuming the cable itself can be built.
I think that's the paper's main weakness, actually: its reliance on finding an epoxy to construct the cable with, that will allow the overall cable strength to be similar to the inherent nanotube strength. The proposal calls for 3-cm lengths of carbon nanotube to be assembled into the cable (in a mostly flat ribbon shape) with the epoxy. This is because such lengths of carbon nanotubes have indeed been produced, and the paper is trying to go with known technology as much as possible.
Now it seems to me that finding an epoxy strong enough to hold on to the fibers would require finding a substance with nearly as much strength as the fibers themselves. Otherwise, the epoxy will fail when the load becomes great, and the fibers will just slip out. A strong rope does you no good if you can't hold onto it!
Though perhaps there's something about epoxying materials from fibers that I don't understand. Anyone? Anyone? Bueller?
The terrorist-threat angle is another concern. Though a terrorist attack would presumably occur at the low end of the cable, which would have minimal effect on the earth.
The main environmental risk is that of the cable breaking at a high point, possibly at the counterweight. The paper say that if this happens, "About 3000 kg of 2 square millimeter cross-section cable... may fall to Earth intact and east of the anchor." It goes onto say that further study/simulation is necessary to determine the full threat.
So again, for me, I'm not so sure that the epoxy technique of cable construction will work. We may have to wait until we have enough nano-scale control to be able to construct the cable with full-cable-length nanotube, finely interwoven. Of course, once we can make nanotubes like that, a lot of other possibilities for space travel may open up.
The one whose velocity will change more is the one with the smaller mass -- and that'd be the base station.
I found the paper
on the space elevator by Bradley C. Edwards. From section 10.8:
Small oscillations or traveling waves that may be induced by wind or meteors can also be actively damped out at the base of the cable if a cable displacement monitoring system is implemented to detect any movements in the cable.
One oscillation that Pearson [in an earlier paper on a somewhat different elevator setup] investigated was that of transverse waves induced by climbers. The bottom line on this oscillation is that large oscillations can be induced when the climber transverses the length of the cable in one period of the cable s characteristic frequency. (Pearson assumed no counterweight so had the climber traveling twice the length of the cable during one period.) Since we just calculated our cable's characteristic period to be 7.1 hours we will only need to worry about this particular affect when we plan to have climbers traveling at close to 10,000 km/hr.
So looks like the details of the scenario will make it a workable problem after all: a given elevator/climber will cause a transverse wave to travel up and down the cable. At the base station on Earth, an active-dampening system can be employed to cancel out any (small) waves that make it to the bottom (that aren't already damped out/overshadowed by the effects of the atmosphere). The wave that travels up will mostly reflect back downwards if the base station at the top is much more massive than the cable at that point, and if there is an abrupt transition from cable to station. The reflected wave could then be damped at the bottom.
Thus, the lateral force will indeed be communicated down to the bottom, and will ultimately come out of the Earth's rotation.
> What you're doing is taking an object that's spinning at 1 earth circumference/day, and moving it out to a point where it is STILL moving at 1 earth circumference/day. It doesn't speed up or slow down at all.
Ok, I'll go ahead and get into some details.
In terms of how many degrees/day, yes, no difference. However, actual acceleration has to do witjh *linear* velocity. I think you get this, but I'll do a quick calc anyways:
On the surface (at the equator), that's 25,000mi/day, or about 1000 mph. At an altitude of 60,000 miles, that works out to 2*3.14*(60000+4000) miles/day, or about 17,000 mph. So that means you need to apply a lateral acceleration to get from 1kmph to 17kmph. If you don't put a rocket or other propulsor on the elevator, that acceleration has to come from the cable.
Now...
> What you have done is increased its angular momentum, since you increased its radius.
Yes, exactly -- more linear speed, more momentum.
> That increase came from the Earth, since the elevator remained taut during the entire time (again, gravity is doing this) and applied a torque against the Earth.
I think this is your mistake -- you're assuming that the elevator's base station up at 60k miles is somehow permanently fixed relative to the earth. It is not. It is in orbit, and if you apply a lateral acceleration to it, it will speed up or slow down just like any other satellite.
Think about it: you have a force being applied laterally on the cable somewhere in the middle. There are 2 larger masses attached to this cable, one on each end: the earth, and the base station. The one whose velocity will change more is the one with the smaller mass -- and that'd be the base station. Yes, you'd affect the earth's angular velocity a little bit too, but because it's so much more massive than the base station, the difference will be negligible relative to the effect on the station.
So, you're going to have to compensate -- either by using propulsion at the base station, or by balancing the lateral forces by keeping the same amount of mass going down as coming up. The balancing would obviously be the preferable solution.
As for your skater analogy, imagine she has two balls on strings instead of having her arms outstretched, while she's spinning. If she pulls in the strings, making the balls come closer, they're going to start rotating ahead of the rest of her body; if she pulls hard enough, the balls will come around and whack her. The reason that doesn't happen with her arms are because her arms are relatively *rigid* -- they are able to communicate these lateral forces back into the main body, allowing the whole system to equalize and spin as a rigid body.
Ok, now on to...
> Air resistance.
I should have described this in my original post: this rail gun/maglev launching device would have to terminate a a very high altitude -- one high enough to be above at least the majority of the atmosphere. But I think this would be no more difficult than building the base station for the elevator -- which, according to at least one proposal I've heard, would have to be something like 10km high. Build that on top of a mountain, you'll get another few kilometers. That may well be high enough to make atmo resistance a relatively minor concern (remember, atmo density decreases exponentially as you go up). Though of course, I don't know the precise numbers.
If this holds up, I would regard that as an equivalent challenge, between the rail and the elevator. Though the rail *would* need something else too -- a tunnel along its entire length, which can be evacuated of all air. I admit, this may be difficult.
> You get to Mars in a matter of a few days by climbing a very large cable, and then letting go.
I'm not so sure those numbers would work out -- if you have a reference, I'd be interested to see it. For now, let me go through a few mental steps:
- If you release *at* the base station, you'll just float along beside. - If you release a *little* ways out, you'll go away from the station, but you won't be at escape velocity. You'll just be in an elliptical orbit, with the point at which you detatched from the outer cable being the closest point (apogee or perigee? I can't remember). - There will be some critical distance where you'll be at escape velocity when you detach. But if you're only just above that, you're really not going to be going that fast. The craft that went to the moon were going only just above escape velocity, and it took several days. Our current mars probes are going a bit faster than that, and take 18 months to arrive.
So then the question is: how far out do you have to go to get the kinds of speeds you're talking about? Will there be enough outward length of cable for this? We'd have to work out the numbers.
Note that lateral acceleration problem comes up again: as you let masses go out along this outer cable, you're going to need to have a balance for the lateral acceleration you'd get. That means you'll have to have other masses arriving at the cable somewhere much farther out, and being drawn in at the same time. Could get tricky.
-----
Ok, so anyways, after all this discussion, I want to say that I regret the title I gave to my first post. The space elevator would definitely have a place in an overall spacefaring economy. But I don't think it'd be something that would be relied upon exclusively, in the long term.
> Okay, I'll simplify everything I'm going to say right now.
I'll excuse that, since my own post was a bit flamish.
>... rocket-type launchers.
I never said anything about rockets. A rail gun uses electomagnetic effects.
If the space elevator uses EM to raise an elevator, it's using roughly the same technology, only it's a track that's maybe 10,000 times as long. Even when you recoup the energy from the descending elevators, the inefficiencies will probably add up to more than the energy required to do a rail-gun launch (into LEO, at least).
And I don't see any reason why you couldn't pipeline a rail gun just as much as a space elevator.
> You're stealing rotational inertia from the Earth to speed things up. Gravity keeps it from slowing down.
I think you're the one who needs to brush up on his physics. Or maybe geometry.
But it doesn't matter -- see the other response to my post, regarding how it's not a problem due to descending elevators cancelling it out -- as long as you have equivalent mass going up/down.
As far as economics, think about this: one of the major reasons why the shuttle program is so expensive is the monolithic nature of a government agency like NASA. If incentives were found for private corporations to invest in manned space flights, the process would soon get much cheaper, technology would advance faster, etc. etc. I refer you to maybe half the other posts in this discussion.
A space elevator will be an inherently monolithic project. It's just begging to be a pork-laden, budget-overunning, gigantic money hole. And think of the power of whatever agency/government/corporation winds up controlling it!
A smaller-scale rail gun or similar design would be much easier to build, and therefore more distributed. Competition would be possible, driving costs down much further.
> Ultimately, we could have dozens of space elevators...
But they'd be aligned in a single ring along the equator, instead of having perhaps thousands of rail guns arrayed across the globe. Which will be easier to use? And which one will ultimately be more bandwith-limiting?
> But an initial space elevator will be a loss leader.
I wonder about that. To build the next elevator, if we're talking about material brought up from the Earth, it could still take a lot more energy/time to do that, even with an existing space elevator, than to build a rail-gun launcher. That's due to the sheer *amount* of material we're talking about for an elevator -- it's not quite so ribbon-thin at the higher altitudes!
If we're talking about capturing an asteroid, putting it in the proper orbit, minining it, etc -- well, that's even *more* effort, isn't it?
> You're ignoring one very important point: cargo goes up, cargo comes down
Hmm, yes, I was. I focused my argument too much on just one side of the equation. But...
> The net energy requirements are tiny compared with launching cargo into space.
I wouldn't say that's a sure thing. You're not going to retrieve 100% of the energy on the way down. Even if you only lose a few percent, over such a vast distance, I could see that very well adding up to more than the energy cost of a rail-gun launch. This could especially be true if you only want low-earth-orbit. I don't know any actual numbers, of course.
Also, you still need a way to go from the top of the elevator to your final destination. True, the elevator-top would itself be an attractive destination, but we're talking the bigger picture here. Anyways, you'd need to keep a supply of propellant of some kind up there. If you have to transport it from Earth, that effectively cuts the efficiency of the elevator. With rail guns, you could very well give your craft enough energy to make it all the way to the moon, mars, etc, with only minimal manuvering thrusters needed on the ship.
Yes, you can just let ships fly away from the elevator-top due to centrifugal effect, but you might still need more additional acceleration to be provided by the ship itself than you would if a rail gun was providing the initial energy.
> Balancing rising and descending loads also addresses your lateral acceleration problem.
Hum, yes. One minor problem might be interference between waves travelling up/down with the elevators. Though careful scheduling of the elevator trips would probably avoid any constructive interference from becoming, er, destructive.
> > Something else to think about: why are automobiles so much more popular than trains? > Automobiles (I assume you're talking about trucks) are less useful than trains when it comes to transporting bulk cargo. They're also less efficient in terms of energy used per weight transported.
But then why are trucks so popular? Because of the freedom gained by having a much larger number of source/destination points. If you're a company in North America needing to get a 2-ton item to the moon, how would the cost analysis work out between:
A) hauling it, what, 5000 miles to the equator, sending it up on the elevator, then transferring it to another vehicle to go to the moon, vs.
B) sending it 100 miles to the nearest rail gun, and launching it directly to the moon?
The rail launch itself may be more expensive, but the total cost (in $$) may be less. And it may be the only option if you need it there quick.
I was also referring to passenger automobiles -- the same logic applies to why cars are so damn much more popular than trains. Yes, they're a lot less energy efficient, and more expensive, but they afford a much greater degree of freedom. People are willing to pay extra for that -- in energy, money, and even risk to life.
Actually, I guess I can see a place for both technologies. As passenger/freight trains/slow boats across the ocean still have a place where low cost outweighs timeliness, the space elevator may have a role. But for most space travel, I see a rail-gun or similar approach being more popular and even economical, if not more energy-efficient.
Quite a few reasons, but the biggest one can be summed up in 2 words: SERIAL CONNECTION.
You'd only be able to lift one batch of cargo at at time. Yes, you could pipeline many of them, but they'd still all have to start at one point on the earth, and all go to a single destination. And any hiccup in one would stop all of them.
How would one lift these cargo loads, anyways? Well, electromagnetic techniques seem reasonable, but what if you took the resources it would require to build that 100,000-kilometer-long EM lift and split it into one thousand 100-km-long rail guns? They'd be able to shoot loads into space, in a variety of directions, each at the same rate as the single elevator. Which means you'd be able to put 1000 times as much stuff into space in the same amount of time. i.e. do it in PARALLEL.
Of course, you'd need a lot more power for each of these rail guns as the equivalent-length section of the elevator. But you know, they probably don't each need to be 100km. Maybe 10km? Also, maybe you don't need a thousand of them. Try 100. So probably for the same power expenditure, and 1/100th the construction materials, you'd be able to deliver freight/passengers into space at 100 times the rate.
One other thing people seem to forget about a space elevator: rotational inertia. As you raise cargo up, you're going to have to accelerate it laterally as well, or your elevator will get pushed backwards. So you'd either have to constantly be using thrusters on the base station at the top, or each load would need a rocket. Which negates the supposed "no dangerous fuel" advantage.
Something else to think about: why are automobiles so much more popular than trains?
> I had always assumed that the front compartment would be highly reinforced...
Unless it had its own smooth, tight covering of those re-entry tiles, it was still going way too fast to survive. We're talking 12,000 mph. It was designed to survive *launch*, where it really wouldn't get up to that type of speed by the time it left the atmosphere.
>... and could remain intact longer than the rest of it. It seems that is not the case.
That said, it probably *did* survive longer -- but not my much. Maybe the experts will be able to pick out which of those pieces streaming through the sky was the cabin.
Yeah, there'd have to be a server to keep track of where it is, and to keep updating it as it gets shifted around -- and if someone cracks into the server...
> > HT can increase performance a lot more than more cache
> I disagree with this, because page faults are are pretty expensive relative to a trip-up in a CPU pipeline.
Though if you have an HT system, the application you're using will only be taking up one half of the cache. Meanwhile, the system is sitting there, keeping hold of the other half (unless you have an actual other thread using it). (That's how the pentium HT works, btw -- the cache is split in half, one for each virtual processor). When you go to do a system task, such as switching windows, you might save a page fault due to the window manager, if all the info needed is sitting in the system's cache.
True, HT won't benefit most individual user applications, but it may make it quicker to switch between them.
I have not seen whether this is true in practice, however.
> What do I do with the two computers I already have?... I can't simply buy the new processors and pop them into the hardware I already have.
But you can plop in a new MoBo + processor (well, ok, not with the laptop, but certainly with the others). You may even be able to re-use the RAM. No need to get a new case/cd-rom/monitor/etc -- which comprise most of the hard-to-get-rid-of junk.
You might even be able to use the bare mobo+processor (+ram) as naked compute servers. You'd have to get power supplies, and ethernet cards if not built into the mobo's, and have a place to put them (some woodworking skills may help). And you'd have to get them to network-boot, etc. I plan on doing this myself... er, someday.:)
> If you have 10 megs to back up, and that needs to be redundantly copied over 100 machines
The original post wasn't referring to having an *entire* copy of your data to 100 machines -- simply that it would be *split up* among 100 machines. Though the parity bits do take up some extra space -- probably wouldn't need nearly the doubling of data that he proposed.
But still, yes, there is a matter of who takes the space vs. who gives it -- if this system is popular enough, the "excess space" will run out!
Slashdot Mirror
From what I hear generally, the main barrier for OLED's is the longevity of the devices. That might be what'll take it a while to get going, because it's still kind of a fundamental engineering problem. Though -- if it's cheap enough, it might not be so bad to junk your monitor after 1 year, and buy another. Though that'd be bad environmentally, etc. Plus it'd be a pain. Ah, well.
Doesn't look like it -- from the diagram, it's looking very directional (the angles of the backplane reflectors are set up to reflect light from the emitter at the center to a line perpendicular to the panel).
One problem: one of the useful things about the cluster-of-bullet-LED's we have now for traffic/brake/etc. lights is an additional fault-tolerance aspect: you could have a substantial number of them malfunctioning and still be useful. I have actually seen this -- a big truck in front of me had a really annoying one, where 1/3 of them were out, 1/3 on, and 1/3 flickering randomly. Annoying, but it still performed its function.
If you just have one LED chip in the middle, if it goes out, you gotta replace the whole thing.
Oh, it also seems that it'll be harder to get sufficient brightness for the really demanding applications (e.g. a traffic light with the sun behind it).
Though for home lighting -- yeah, baby!
You mean when you can't read those -- nothing to do with availablility of drives, they're going to rot!. Better get them onto a cd as soon as you can.
I've still got a lot of pr0n^H^H^H^H Data on cd. What will I do if I can't read those?
You mean when you can't read those -- nothing to do with availablility of drives, they're going to rot!. Better get them onto a dvd as soon as you can.
I've still got a lot of pr0n^H^H^H^H Home videos on dvd. What will I do if I can't read those?
You mean when you can't read those -- nothing to do with availablility of drives, they're going to rot!. Better get them onto a holo-crystal as soon as you can.
I've still got a lot of love-slave android personaliti^H^H^H^H Personal memory engram backups on holo-crystal. What will I do if I can't read those?
Ah, but in Soviet Russia, what will you do when your holo-crystal can't read YOU!
I'd tend to think that would fall under the category of "if you could do that, you'd already be able to build a robot to do it."
Or or those who don't like to exercise, have the computer work out while you surf the web.
Now that I like the sound of!
For work: You can visualize nice models in space.
Ok, I'm thinking Cindy Crawford -- but then it's not so great with the space suit in the way.
What does that have to do with work, by the way?
I believe he was referring to the matter of free speech.
Are you saying it is RIGHT to boycott ...[various examples]
I think he was not saying what is right, but more of what one has a right to do.
Doesn't matter. One has the right to boycott what one wishes. If it is an unfair/unreasonable boycott, then additional voices can be raised against it.
Hey, that'd be a meta-boycott!
Imagine what a beowulf cluster of meta boycotts could... ouch!
"The zoo? You believe what the zoo tells you?"
Yes, indeed, programmatic shaders are one of the most important features of the Renerman standard. But I'm betting that, at Pixar, 90% of the time they use texmaps. It just works out to be a lot easier to do it that way -- provided you have an army of artists to generate all those 2D images! (BTW, technically, renderman always uses programmatic shaders; to do texmaps, you use a simple shader that just goes and looks up the color values to set.)
Plus, since it typically isn't doing raytraycing, it needs to do a lot of "shadow map" lookups, to tell whether an object is being directly lit or not. I don't know all the details of that, but it does involve doing a texture-map-like lookup for every light source.
Thanks for the "Renderman" etymology. :-)
You're welcome! I got it from the Advanced Renderman book. There's a few paragraphs on it in the back.
What? They're finding porn in nebulas now?
Oh, wait... darn
Aye, there's the rub.
Given that they picked Intel chips over Athlons, and given that they must have carefully compared chip performance on their particular application (i.e. Renderman), that says to me that Renderman is mostly memory-bandwith limited, rather than heavy-math-calculation limited.
I know a little bit about how that program works: it's not a ray-tracer. It does some basic 3d calcs to trace lines from camera to objects, and to subdivide polygons. But to determine the actual color of each pixel, it's mostly a matter of one or a number of texture-map, shadow-map, and other lookups. Each is addressing a small part of a big range of memory -- probably breaks through the cache incessantly. I'd bet that many of the geometric calculations are memory-limited too, due to the absolutely humongous number of objects they put into a given scene (e.g. blue fur).
From what I understand, DSP's are good for 2D image processing -- because the algorithms are fairly standard, and require a lot of signal-processing-like math. For 3D, perhaps the matrix calcs involved in the coordinate transforms could be done in a dsp, but as I described, that's probably not a big enough piece of the puzzle to make it worth it.
To go a little further: dedicated hardware was actually the original goal of Pixar, even before it was split off as its own company. But they noticed that hardware advances were so fast that their designs were getting obsolete fast. (The name "renderman" actually came from a quip by one of their engineers, commenting on how they'd soon be able to design a machine that could fit in a pocket-sized device that you could carry around like a Walkman.) Anyways, they eventually discarded the custom hardware, because their software-only "practice" version was getting quite acceptable performance levels all by itself -- on general-purpose processors.
I think you've missed something here: what about the excellent astronaut-training facilities/staff/experience that NASA has built up over the decades, as well as the current crop of astronauts? I'd suggest that in addition to your "Fifth" option, add the creation of NATA, the National Astronautics Training Academy. This would consist of just that subset of NASA; its job would be to train the astronauts that companies would hire for space work.
A company may be able to finance the hardware for a space mission, but the astronaut training requires a big, big, long-term investment. Keeping that centralized probably makes sense. This would be a sort of "rent-a-naut" system. :)
Here's a good paper on the subject. It's a 15mb pdf, but worth the download.
In it, a good many of the technical problems are solidly examined, and reasonable solutions are proposed.
The approach presented is to launch an initial spool of very thin cable into geosynchronous orbit. This spool will be some thousands of kg in mass; this won't be *that* much harder than putting up a communications satellite. Then you lower that cable down to earth (and raise spool-unreeling spacecraft up past geosynch. as a counterweight), and you have a sort of "mini" space elevator that can haul up a mere 1200 kg. A series of climbers then ascend, each epoxying on a new layer of cable. Continue for 2 years, and you have a cable that can carry up as much as the shuttle. Continue for 5 years, and you have one that can lift a million kg.
All the solutions to the technical problems will require lots of research/testing to truly overcome, so it'd likely still be decades away, even with full effort. And that's also assuming the cable itself can be built.
I think that's the paper's main weakness, actually: its reliance on finding an epoxy to construct the cable with, that will allow the overall cable strength to be similar to the inherent nanotube strength. The proposal calls for 3-cm lengths of carbon nanotube to be assembled into the cable (in a mostly flat ribbon shape) with the epoxy. This is because such lengths of carbon nanotubes have indeed been produced, and the paper is trying to go with known technology as much as possible.
Now it seems to me that finding an epoxy strong enough to hold on to the fibers would require finding a substance with nearly as much strength as the fibers themselves. Otherwise, the epoxy will fail when the load becomes great, and the fibers will just slip out. A strong rope does you no good if you can't hold onto it!
Though perhaps there's something about epoxying materials from fibers that I don't understand. Anyone? Anyone? Bueller?
The terrorist-threat angle is another concern. Though a terrorist attack would presumably occur at the low end of the cable, which would have minimal effect on the earth.
The main environmental risk is that of the cable breaking at a high point, possibly at the counterweight. The paper say that if this happens, "About 3000 kg of 2 square millimeter cross-section cable ... may fall to Earth intact and east of the anchor." It goes onto say that further study/simulation is necessary to determine the full threat.
So again, for me, I'm not so sure that the epoxy technique of cable construction will work. We may have to wait until we have enough nano-scale control to be able to construct the cable with full-cable-length nanotube, finely interwoven. Of course, once we can make nanotubes like that, a lot of other possibilities for space travel may open up.
I found the paper on the space elevator by Bradley C. Edwards. From section 10.8:
So looks like the details of the scenario will make it a workable problem after all: a given elevator/climber will cause a transverse wave to travel up and down the cable. At the base station on Earth, an active-dampening system can be employed to cancel out any (small) waves that make it to the bottom (that aren't already damped out/overshadowed by the effects of the atmosphere). The wave that travels up will mostly reflect back downwards if the base station at the top is much more massive than the cable at that point, and if there is an abrupt transition from cable to station. The reflected wave could then be damped at the bottom.Thus, the lateral force will indeed be communicated down to the bottom, and will ultimately come out of the Earth's rotation.
> What you're doing is taking an object that's spinning at 1 earth circumference/day, and moving it out to a point where it is STILL moving at 1 earth circumference/day. It doesn't speed up or slow down at all.
Ok, I'll go ahead and get into some details.
In terms of how many degrees/day, yes, no difference. However, actual acceleration has to do witjh *linear* velocity. I think you get this, but I'll do a quick calc anyways:
On the surface (at the equator), that's 25,000mi/day, or about 1000 mph. At an altitude of 60,000 miles, that works out to 2*3.14*(60000+4000) miles/day, or about 17,000 mph. So that means you need to apply a lateral acceleration to get from 1kmph to 17kmph. If you don't put a rocket or other propulsor on the elevator, that acceleration has to come from the cable.
Now...
> What you have done is increased its angular momentum, since you increased its radius.
Yes, exactly -- more linear speed, more momentum.
> That increase came from the Earth, since the elevator remained taut during the entire time (again, gravity is doing this) and applied a torque against the Earth.
I think this is your mistake -- you're assuming that the elevator's base station up at 60k miles is somehow permanently fixed relative to the earth. It is not. It is in orbit, and if you apply a lateral acceleration to it, it will speed up or slow down just like any other satellite.
Think about it: you have a force being applied laterally on the cable somewhere in the middle. There are 2 larger masses attached to this cable, one on each end: the earth, and the base station. The one whose velocity will change more is the one with the smaller mass -- and that'd be the base station. Yes, you'd affect the earth's angular velocity a little bit too, but because it's so much more massive than the base station, the difference will be negligible relative to the effect on the station.
So, you're going to have to compensate -- either by using propulsion at the base station, or by balancing the lateral forces by keeping the same amount of mass going down as coming up. The balancing would obviously be the preferable solution.
As for your skater analogy, imagine she has two balls on strings instead of having her arms outstretched, while she's spinning. If she pulls in the strings, making the balls come closer, they're going to start rotating ahead of the rest of her body; if she pulls hard enough, the balls will come around and whack her. The reason that doesn't happen with her arms are because her arms are relatively *rigid* -- they are able to communicate these lateral forces back into the main body, allowing the whole system to equalize and spin as a rigid body.
Ok, now on to...
> Air resistance.
I should have described this in my original post: this rail gun/maglev launching device would have to terminate a a very high altitude -- one high enough to be above at least the majority of the atmosphere. But I think this would be no more difficult than building the base station for the elevator -- which, according to at least one proposal I've heard, would have to be something like 10km high. Build that on top of a mountain, you'll get another few kilometers. That may well be high enough to make atmo resistance a relatively minor concern (remember, atmo density decreases exponentially as you go up). Though of course, I don't know the precise numbers.
If this holds up, I would regard that as an equivalent challenge, between the rail and the elevator. Though the rail *would* need something else too -- a tunnel along its entire length, which can be evacuated of all air. I admit, this may be difficult.
> You get to Mars in a matter of a few days by climbing a very large cable, and then letting go.
I'm not so sure those numbers would work out -- if you have a reference, I'd be interested to see it. For now, let me go through a few mental steps:
- If you release *at* the base station, you'll just float along beside.
- If you release a *little* ways out, you'll go away from the station, but you won't be at escape velocity. You'll just be in an elliptical orbit, with the point at which you detatched from the outer cable being the closest point (apogee or perigee? I can't remember).
- There will be some critical distance where you'll be at escape velocity when you detach. But if you're only just above that, you're really not going to be going that fast. The craft that went to the moon were going only just above escape velocity, and it took several days. Our current mars probes are going a bit faster than that, and take 18 months to arrive.
So then the question is: how far out do you have to go to get the kinds of speeds you're talking about? Will there be enough outward length of cable for this? We'd have to work out the numbers.
Note that lateral acceleration problem comes up again: as you let masses go out along this outer cable, you're going to need to have a balance for the lateral acceleration you'd get. That means you'll have to have other masses arriving at the cable somewhere much farther out, and being drawn in at the same time. Could get tricky.
-----
Ok, so anyways, after all this discussion, I want to say that I regret the title I gave to my first post. The space elevator would definitely have a place in an overall spacefaring economy. But I don't think it'd be something that would be relied upon exclusively, in the long term.
> If you're wondering why you didn't get modded up, it's because it is entirely possible (and probable) to have multiple elevators on the same cable.
I guess you didn't see the word "pipeline" I used.
> Okay, I'll simplify everything I'm going to say right now.
... rocket-type launchers.
I'll excuse that, since my own post was a bit flamish.
>
I never said anything about rockets. A rail gun uses electomagnetic effects.
If the space elevator uses EM to raise an elevator, it's using roughly the same technology, only it's a track that's maybe 10,000 times as long. Even when you recoup the energy from the descending elevators, the inefficiencies will probably add up to more than the energy required to do a rail-gun launch (into LEO, at least).
And I don't see any reason why you couldn't pipeline a rail gun just as much as a space elevator.
> You're stealing rotational inertia from the Earth to speed things up. Gravity keeps it from slowing down.
I think you're the one who needs to brush up on his physics. Or maybe geometry.
But it doesn't matter -- see the other response to my post, regarding how it's not a problem due to descending elevators cancelling it out -- as long as you have equivalent mass going up/down.
As far as economics, think about this: one of the major reasons why the shuttle program is so expensive is the monolithic nature of a government agency like NASA. If incentives were found for private corporations to invest in manned space flights, the process would soon get much cheaper, technology would advance faster, etc. etc. I refer you to maybe half the other posts in this discussion.
A space elevator will be an inherently monolithic project. It's just begging to be a pork-laden, budget-overunning, gigantic money hole. And think of the power of whatever agency/government/corporation winds up controlling it!
A smaller-scale rail gun or similar design would be much easier to build, and therefore more distributed. Competition would be possible, driving costs down much further.
> Ultimately, we could have dozens of space elevators...
But they'd be aligned in a single ring along the equator, instead of having perhaps thousands of rail guns arrayed across the globe. Which will be easier to use? And which one will ultimately be more bandwith-limiting?
> But an initial space elevator will be a loss leader.
I wonder about that. To build the next elevator, if we're talking about material brought up from the Earth, it could still take a lot more energy/time to do that, even with an existing space elevator, than to build a rail-gun launcher. That's due to the sheer *amount* of material we're talking about for an elevator -- it's not quite so ribbon-thin at the higher altitudes!
If we're talking about capturing an asteroid, putting it in the proper orbit, minining it, etc -- well, that's even *more* effort, isn't it?
> You're ignoring one very important point: cargo goes up, cargo comes down
Hmm, yes, I was. I focused my argument too much on just one side of the equation. But...
> The net energy requirements are tiny compared with launching cargo into space.
I wouldn't say that's a sure thing. You're not going to retrieve 100% of the energy on the way down. Even if you only lose a few percent, over such a vast distance, I could see that very well adding up to more than the energy cost of a rail-gun launch. This could especially be true if you only want low-earth-orbit. I don't know any actual numbers, of course.
Also, you still need a way to go from the top of the elevator to your final destination. True, the elevator-top would itself be an attractive destination, but we're talking the bigger picture here. Anyways, you'd need to keep a supply of propellant of some kind up there. If you have to transport it from Earth, that effectively cuts the efficiency of the elevator. With rail guns, you could very well give your craft enough energy to make it all the way to the moon, mars, etc, with only minimal manuvering thrusters needed on the ship.
Yes, you can just let ships fly away from the elevator-top due to centrifugal effect, but you might still need more additional acceleration to be provided by the ship itself than you would if a rail gun was providing the initial energy.
> Balancing rising and descending loads also addresses your lateral acceleration problem.
Hum, yes. One minor problem might be interference between waves travelling up/down with the elevators. Though careful scheduling of the elevator trips would probably avoid any constructive interference from becoming, er, destructive.
> > Something else to think about: why are automobiles so much more popular than trains?
> Automobiles (I assume you're talking about trucks) are less useful than trains when it comes to transporting bulk cargo. They're also less efficient in terms of energy used per weight transported.
But then why are trucks so popular? Because of the freedom gained by having a much larger number of source/destination points. If you're a company in North America needing to get a 2-ton item to the moon, how would the cost analysis work out between:
A) hauling it, what, 5000 miles to the equator, sending it up on the elevator, then transferring it to another vehicle to go to the moon, vs.
B) sending it 100 miles to the nearest rail gun, and launching it directly to the moon?
The rail launch itself may be more expensive, but the total cost (in $$) may be less. And it may be the only option if you need it there quick.
I was also referring to passenger automobiles -- the same logic applies to why cars are so damn much more popular than trains. Yes, they're a lot less energy efficient, and more expensive, but they afford a much greater degree of freedom. People are willing to pay extra for that -- in energy, money, and even risk to life.
Actually, I guess I can see a place for both technologies. As passenger/freight trains/slow boats across the ocean still have a place where low cost outweighs timeliness, the space elevator may have a role. But for most space travel, I see a rail-gun or similar approach being more popular and even economical, if not more energy-efficient.
Quite a few reasons, but the biggest one can be summed up in 2 words: SERIAL CONNECTION.
You'd only be able to lift one batch of cargo at at time. Yes, you could pipeline many of them, but they'd still all have to start at one point on the earth, and all go to a single destination. And any hiccup in one would stop all of them.
How would one lift these cargo loads, anyways? Well, electromagnetic techniques seem reasonable, but what if you took the resources it would require to build that 100,000-kilometer-long EM lift and split it into one thousand 100-km-long rail guns? They'd be able to shoot loads into space, in a variety of directions, each at the same rate as the single elevator. Which means you'd be able to put 1000 times as much stuff into space in the same amount of time. i.e. do it in PARALLEL.
Of course, you'd need a lot more power for each of these rail guns as the equivalent-length section of the elevator. But you know, they probably don't each need to be 100km. Maybe 10km? Also, maybe you don't need a thousand of them. Try 100. So probably for the same power expenditure, and 1/100th the construction materials, you'd be able to deliver freight/passengers into space at 100 times the rate.
One other thing people seem to forget about a space elevator: rotational inertia. As you raise cargo up, you're going to have to accelerate it laterally as well, or your elevator will get pushed backwards. So you'd either have to constantly be using thrusters on the base station at the top, or each load would need a rocket. Which negates the supposed "no dangerous fuel" advantage.
Something else to think about: why are automobiles so much more popular than trains?
Ok. Done ranting for now.
> > Of course, the reason they want to kill you has nothing to do with their "tan colored skin" (and everything to do with their f*cked-up religion)
> Oh, now I get it! We're supposed to be discriminating against Muslims, not just middle-eastern people!
I'd tend to think the "religion" being referred to is not Islam in general, but that bastardized version that the Al Quaeda folks adhere to.
> I had always assumed that the front compartment would be highly reinforced...
... and could remain intact longer than the rest of it. It seems that is not the case.
Unless it had its own smooth, tight covering of those re-entry tiles, it was still going way too fast to survive. We're talking 12,000 mph. It was designed
to survive *launch*, where it really wouldn't get up to that type of speed by the time it left the atmosphere.
>
That said, it probably *did* survive longer -- but not my much. Maybe the experts will be able to pick out which of those pieces streaming through the sky was the cabin.
*sigh*
Yeah, there'd have to be a server to keep track of where it is, and to keep updating it as it gets shifted around -- and if someone cracks into the server...
> HyperThreading (as implemented by Intel...remember they didn't invent this idea) shares a single cache line among two virtual processors.
Nope, the cache is actually split in half. Each virtual processor can access only its half.
> This is also why when there are HT problems, the speed drop can be terrible.
The main speed drop would be due to the smaller available cache size (per thread).
> > HT can increase performance a lot more than more cache
> I disagree with this, because page faults are are pretty expensive relative to a trip-up in a CPU pipeline.
Though if you have an HT system, the application you're using will only be taking up one half of the cache. Meanwhile, the system is sitting there, keeping hold of the other half (unless you have an actual other thread using it). (That's how the pentium HT works, btw -- the cache is split in half, one for each virtual processor). When you go to do a system task, such as switching windows, you might save a page fault due to the window manager, if all the info needed is sitting in the system's cache.
True, HT won't benefit most individual user applications, but it may make it quicker to switch between them.
I have not seen whether this is true in practice, however.
> What do I do with the two computers I already have? ... I can't simply buy the new processors and pop them into the hardware I already have.
:)
But you can plop in a new MoBo + processor (well, ok, not with the laptop, but certainly with the others). You may even be able to re-use the RAM. No need to get a new case/cd-rom/monitor/etc -- which comprise most of the hard-to-get-rid-of junk.
You might even be able to use the bare mobo+processor (+ram) as naked compute servers. You'd have to get power supplies, and ethernet cards if not built into the mobo's, and have a place to put them (some woodworking skills may help). And you'd have to get them to network-boot, etc. I plan on doing this myself... er, someday.
> If you have 10 megs to back up, and that needs to be redundantly copied over 100 machines
The original post wasn't referring to having an *entire* copy of your data to 100 machines -- simply that it would be *split up* among 100 machines. Though the parity bits do take up some extra space -- probably wouldn't need nearly the doubling of data that he proposed.
But still, yes, there is a matter of who takes the space vs. who gives it -- if this system is popular enough, the "excess space" will run out!